Colon Cancer Blood Test

Colorectal cancer (CRC) is the second most common form of cancer worldwide. While screening programs have been shown to reduce mortality by up to 30 percent, many individuals don’t follow through with testing because of the pain and expense associated with colonoscopy. Now, a less invasive, data-driven screening tool could one day replace colonoscopies for standard CRC screening. Called ColonFlag, the system uses machine learning to analyze a blood sample, and has been demonstrated to be twice as accurate as a fecal exam.

ColonFlag compares current samples to a patient’s historical profile as well as a database of 20 million anonymized tests. The algorithms can automatically detect subtle changes and correlations to predict the likelihood of developing CRC earlier than traditional methods.

A blind study is currently underway to quantify the efficacy of the new method. The developer is also working with two major U.S. hospitals to find other potential applications for their predictive analytic technique, including diabetes, heart failure, kidney injury and sepsis.